A new improved simulated annealing for traveling salesman problem

Abstract

Simulated annealing algorithm is one of the most popular metaheuristics that has been successfully applied to many optimization problems. The main advantage of SA is its ability to escape from local optima by allowing hill-climbing moves and exploring new solutions at the beginning of the search process. One of its drawbacks is its slow convergence, requiring high computational time with a good set of parameter values to find a reasonable solution. In this work, a new improved SA is proposed to solve the well-known travelling salesman problem. In order to improve SA performance, a population-based improvement procedure is incorporated after the acceptance phase of SA, allowing the algorithm to take advantage of the social behavior of some solutions from the search space. Numerical results were carried out using known TSP instances from TSPLIB and preliminary results show that the proposed algorithm outperforms in terms of solution quality, the other comparison algorithms.